Ambari HDP集群搭建文档

一、配置主机和节点机器之间SSH无密登录

多台外网服务器配置时，需要在/etc/hosts中把本机的IP地址设置为内网IP地址

http://2d67df38.wiz02.com/share/s/0JpZYU2X7kl626aE730Q_Flq3daBjm08rAnl2oLpli2uMOfo

二、为每台主机关闭iptables,SELinux,启动ntpd服务

三、安装Mysql,JDK

查看linux发行版本

cat /etc/redhat-release

从网站http://dev.mysql.com/downloads/repo/yum/下载对应版本yum资源

安装

yum -y install mysql57-community-release-el7-7.noarch.rpm

安装mysql数据库的服务器版本

yum -y install mysql-community-server

 

启动数据库

service  mysqld start

systemctl start  mysqld.service

 

获取初始密码

#######从mysql进程日志中获取root用户的初始密码：ra%yk7urCBIh

[root@typecodes ~]# grep "password" /var/log/mysqld.log

 修改root用户密码

mysql -u root -p

 ALTER USER 'root'@'localhost' IDENTIFIED BY 'new password';

 

这个其实与validate_password_policy的值有关。

validate_password_policy有以下取值：

Policy	Tests Performed
0 or LOW	Length
1 or MEDIUM	Length; numeric, lowercase/uppercase, and special characters
2 or STRONG	Length; numeric, lowercase/uppercase, and special characters; dictionary file

首先，修改validate_password_policy参数的值

set global validate_password_policy=0;

这样，判断密码的标准就基于密码的长度了。这个由validate_password_length参数来决定。

validate_password_length参数默认为8

set global validate_password_length=4;

 

安装JDK，配置环境变量

export JAVA_HOME=/usr/local/jdk1.8.0

export JRE_HOME=${JAVA_HOME}/jre

export CLASSPATH=.:${JAVA_HOME}/lib:${JRE_HOME}/lib

export PATH=${JAVA_HOME}/bin:$PATH

使环境变量立即生效

source ~/.bash_profile

四、集群服务器间配置时间同步

http://2d67df38.fromwiz.com/share/s/0JpZYU2X7kl626aE730Q_Flq2OyZYU3BDQcP2rnK4F3zASlT

五、安装Ambari

wget -P /etc/yum.repos.d/ http://public-repo-1.hortonworks.com/ambari/centos6/2.x/updates/2.0.1/ambari.repo

 

wget -P /etc/yum.repos.d/ http://public-repo-1.hortonworks.com/ambari/centos6/2.x/updates/2.2.2.0/ambari.repo

 

yum clean all

 

yum list|grep ambari

yum安装ambari服务端 

yum install ambari-server

 

安装完后进行设置

ambari-server setup

 

 

 到这一步需要安装mysql jar包

yum install mysql-connector-java*

 启动服务

 报错：

 解决方法：

grant all privileges on *.* to ambari@'localhost' identified by 'bigdata';

flush privileges;

 

 重新启动，报错：

创建数据库

create database ambari character set utf8;

 重新启动，报错：

 如果使用自己定义的数据库，必须在启动Ambari服务之前导入Ambari的sql脚本

用Ambari用户（上面设置的用户）登录mysql
mysql -u ambari -p
use ambari
source /var/lib/ambari-server/resources/Ambari-DDL-MySQL-CREATE.sql

启动成功

 日志无报错

 打开 http://localhost:8080

admin/admin

 登录系统操作后发现还是报错：

 21 Feb 2017 19:15:29,519  WARN [com.mchange.v2.async.ThreadPoolAsynchronousRunner$PoolThread-#0] StatementUtils:48 - Statement close FAILED.

com.mysql.jdbc.exceptions.jdbc4.MySQLSyntaxErrorException: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'OPTION SQL_SELECT_LIMIT=DEFAULT' at line 1

原因：Ambari默认的 mysql jdbc 驱动不支持 5.6以上版本

解决方法：下载新的 mysql jdbc驱动

ln -s mysql-connector-java-5.1.40-bin.jar mysql-connector-java.jar

ambari-server setup --jdbc-db=mysql --jdbc-driver=/usr/share/java/mysql-connector-java.jar

 

配置主机时报错

Permanently added 'node1.hadoop' (ECDSA) to the list of known hosts

解决方法

在 master主机用户根目录的.ssh/ssh_config 文件中新增一行

UserKnownHostsFile ~/.ssh/known_hosts

 

继续报错

 解决方法：

修改主机名

vi /etc/sysconfig/network

 vim /etc/hosts

 hostname master.hadoop

 

或者直接 hostnamectl set-hostname 新主机名

 

 

 注意HOST名字要写全名

node2节点报错：

ERROR: Bootstrap of host node2.hadoop fails because previous action finished with non-zero exit code (255)
ERROR MESSAGE: tcgetattr: Invalid argument
Connection to node2.hadoop closed.

STDOUT: tput: unknown terminal "unknown"
('', None)
('', None)

登录到node2上手动启动agent

wget -nv http://public-repo-1.hortonworks.com/ambari/centos7/2.x/updates/2.2.1.0/ambari.repo -O /etc/yum.repos.d/ambari.repo

 

yum repolist

 

yum install ambari-agent

 

vi /etc/ambari-agent/conf/ambari-agent.ini

[server]

hostname=<your.ambari.server.hostname>

url_port=8440

secured_url_port=8441

 

ambari-agent start


 

看后台日志，报错

解决方案：

pip install flask -i http://pypi.douban.com/simple --trusted-host pypi.douban.com

 

删除所有ambari* from /usr/lib/python2.6/site-packages，重新安装ambari-agent,杀死ambari进程，重新启动

rm -f /usr/lib/python2.6/site-packages/ambari_commons
rm -rf /usr/lib/python2.6/site-packages/resource_management
rm -rf /usr/lib/python2.6/site-packages/ambari_jinja2
rm -rf /var/lib/ambari-agent/
rm -rf /usr/lib/ambari-agent/
yum remove ambari-agent

yum install ambari-agent

 

后面又报错：

 

 

 

 Error: Cannot retrieve repository metadata (repomd.xml) for repository: HDP-2.3. Please verify its path and try again

 

解决方法1：

（此方法未验证）

Resolved: The problem was related to proxy Authentication. So, I installed and configured CNTLM to access ISA Proxy Server on Centos 6.

Look at the article :

http://wingloon.com/2014/04/11/how-to-install-and-configure-cntlm-to-access-isa-server-on-centos-6-5/

1.Install and configure cntlm

2. configure yum to use through ISA Proxy

• vi /etc/yum.conf

add a line:

proxy=http://127.0.0.1:3128/

• vi .bash_profile

add :

export http_proxy=http://localhost:3128/

export https_proxy=${http_proxy}

export ftp_proxy=${http_proxy}

• logoff root and login again
• yum clean all
• yum install XXXX

yum update, yum repolist or yum install now works fine

来源： https://community.hortonworks.com/questions/15506/error-cannot-retrieve-repository-metadata-repomdxm.html

解决方法2：

在对应报错服务器上的HDP.repo文件中添加一行  proxy = _none_

 

I have fixed it myself. For those who might face the same issue. This is what I did. Basically HDP.repo should be told to skip proxy server, but if you edit HDP.repo, Ambari wizard will overwrite it. So what is the workaround?. This is how you make HDP.repo skip the proxy. Just add the following lines to the yum.conf file for all the nodes. It worked for me. Hope this helps! :)

[HDP-2.4]

name=HDP-2.4

proxy = _none_

baseurl=http://item-70288/hdp/HDP/centos6/2.x/updates/2.4.0.0

enabled=1

gpgcheck=0

来源： https://community.hortonworks.com/questions/26872/forbidden-403-error-on-hdp-24-installation.html

继续安装，报错：

 解决方法1（本人测试无效，后来用方法2到节点服务器上手工安装）：

在SERVER服务器上执行

vim /etc/ambari-server/conf/ambari.properties

 把标红值改大

解决方案： 1.手动安装yarn apt-get update apt-get install hadoop-2-4-.*-yarn

2.修改配置文件 Can be solved by setting the timeout (agent.package.install.task.timeout=1800) in /etc/ambari-server/conf/ambari.properties

注意： 在安装mapreduce，spark等软件包时，也可能出现类似的问题，请将timeout设置到合适的时间。或者根据错误提示，手动安装相应的软件包。

来源： https://jacobustczhi.gitbooks.io/ambari-ubuntu/chapter5.html

解决方法2：

手工安装

报错:there are unfinished transactions remaining. You might consider running yum-complete-transaction first to finish them.

 

yum update

yum install yum-utils

yum-complete-transaction --cleanup-only

 

报错：

 版本不一致导致，可参考本文最后ambari升级部分

 

后续安装hadoop各组件一般只会是yum超时错误，可手动到服务器上运行yum命令安装，或者下载相关包，制作本地yum源

 

解决不了报错信息，就手动到节点服务器安装agent

http://docs.hortonworks.com/HDPDocuments/Ambari-2.2.1.0/bk_ambari_reference_guide/content/_install_the_ambari_agents_manually.html

rh7

 

wget -nv http://public-repo-1.hortonworks.com/ambari/centos7/2.x/updates/2.2.1.0/ambari.repo -O /etc/yum.repos.d/ambari.repo

 

yum repolist

 

vi /etc/ambari-agent/conf/ambari-agent.ini

将红框处修改为ambari-server主机名

 

ambari-agent start

 

 

ZOOKEEPER安装报错

解决方案：找到zookeeper配置文件，修改服务端配置个数，改为奇数个，

ambari上只需要删除一台节点的zookeeper服务

 

 

PS:在安装某些组件时候，会报配置参数找不到错误

如：

 

可执行如下命令来手动设置参数

/var/lib/ambari-server/resources/scripts/configs.sh -u admin -p admin set 127.0.0.1 bigdata  hive-site hive_server_host "node3.hadoop"

-u ambari系统登录名

-p ambari系统登录密码

bigdata是ambari集群名称

hive_server_host是报错的配置参数名

""内是要设置的配置参数值

 

附：

ambari升级

http://2d67df38.fromwiz.com/share/s/0JpZYU2X7kl626aE730Q_Flq3bDNPQ03b4rH2DzrHN2XmjXh